About — The White Box Project Florida Beta

Contact

Reach out with questions, corrections, additional evidence, data concerns, or any other information relevant to a record shown on this site.

If you believe a record shown here is incorrect, or represents someone whose record has since been sealed, expunged, or otherwise corrected at the source, contact the address above. Records are re-reviewed and, where warranted, corrected or removed.

What This Project Is

The White Box Project audits how race is recorded in publicly available police booking records, using human review and available evidence to identify cases where race has been recorded incorrectly.

Why It Exists

Booking data is used by researchers, journalists, and public databases. In recent years, there have also been increasing examples online of individuals being misidentified as White in booking records despite evidence suggesting otherwise. The White Box Project was created to systematically document these cases. If a race field is wrong at the source, that error will propagate. This project records specific, evidenced examples so the underlying data can be understood more accurately.

What The Florida Beta Includes

The Florida Beta covers Broward, Hillsborough, Orange, Lee, and Polk counties—the five largest counties in Florida for which the project identified readily available booking records that included both a recorded race field and a publicly available booking photograph. The selection began with Florida’s largest counties by population. Miami-Dade County was excluded because comparable recent booking records were not readily available; Palm Beach County because the available records did not include race; and Duval and Pinellas counties because the available records did not include booking photographs. Based on 2025 U.S. Census Bureau population estimates, those exclusions leave Broward, Hillsborough, Orange, Lee, and Polk as the five largest remaining counties meeting the project’s data requirements. For each county, the beta reviews arrests from an approximately 30-day period preceding the completion of that county’s dataset. Because each county was reviewed at a different time, the exact date range varies by county. The applicable period and dataset information for each county are available on the Data page.

Definitions

Confirmed Error: A record for which the available public evidence and human review supports a determination that the listed racial classification is incorrect under the project's review methodology.

Likely Error: A record for which the available public evidence and human review strongly suggests that the listed racial classification is incorrect, but the evidence does not meet the project's threshold for confirmed classification.

Unflagged: A record marked White that has not been classified as either a confirmed or likely error.

Confirmed errors form the confirmed minimum error rate. Confirmed plus likely errors form a maximum possible error rate.

How The Percentages Are Calculated

The denominator for both error rates is the number of records marked White.

Confirmed minimum error rate = confirmed errors ÷ records marked White

Maximum possible error rate = (confirmed errors + likely errors) ÷ records marked White

Percentage marked White = records marked White ÷ total records reviewed

Combined figures across counties are calculated from combined raw totals, not by averaging each county’s own percentage.